The invention relates to the technical field of
data processing, in particular to a
portal vein hypertension monitoring method based on multi-
modal liver
blood vessel parameters, which comprises the following steps: acquiring a CT image, an
MRI image, an ultrasonic image and a liver
pathological section of a patient, and constructing a multi-
modal data source; extracting anatomical structure parameters, functional characteristic parameters and pathophysiologic
phenotype parameters of liver blood vessels from the constructed multi-
modal data source, and constructing a multi-modal representation
library; and inputting the parameters in the multi-modal representation
library into a constructed
hybrid expert model comprising an expert network, a gating network and an integrated mechanism to obtain a predicted value of the hepatic
vein pressure gradient HVPG. Therefore, accurate extraction of
blood vessel parameters under a multi-mode image is used, the anti-
noise capability is improved, and the limitation of a single mode is overcome; and non-invasive detection is performed on the hepatic
vein pressure by using a
hybrid expert model, so that the problem of HVPG invasive detection is solved.